Using artificial intelligence to improve heart stress testing

Researchers are testing whether artificial intelligence and computer analysis can make stress echocardiograms—ultrasound tests of your heart during controlled stress—more accurate at detecting heart disease. Currently, doctors rely on visual interpretation of these images, which can be imprecise. This study uses machine learning to analyze the ultrasound data more objectively and may help doctors diagnose heart problems more reliably in patients presenting with chest pain.

Many people with chest pain need testing to see if they have blocked arteries in their heart, but current stress echocardiogram tests aren't always reliable—about 40–50% of patients who undergo further testing don't actually need it. Computer-assisted analysis of these ultrasound images may help doctors make more accurate diagnoses and avoid unnecessary procedures.